PRODUCTIVITY EFFECTS OF LAND REFORM:

A Study of Disaggregated Farm Data in West Bengal, India 1

Pranab Bardhan2 and Dilip Mookherjee3

September 15, 2008

Abstract

We study the impact of reforms in land property rights on farm productivity in the Indian state

of West Bengal spanning 1982–95. Our analysis is distinguished from earlier studies by use of a

disaggregated farm panel, controls for farm extension programs administered by local governments,

and for endogeneity of program implementation. We find statistically significant effects of tenancy

registration on productivity in tenant farms, but larger general equilibrium spillover effects on non-

tenant farms. The productivity effects of tenancy reform were overshadowed by farm input services

and infrastructure spending by local governments.

1For research support we are grateful to the MacArthur Foundation Inequality Network and the National

Science Foundation (Grant No. SES-0418434). Monica Parra Torrado and Neha Kumar provided outstand-

ing research assistance. We thank officials of the West Bengal government who granted access to the data;

to Sankar Bhaumik and Sukanta Bhattacharya of the Department of Economics, Calcutta University who

led the village survey teams; to Bhaswar Moitra and Biswajeet Chatterjee of the Department of Economics,

Jadavpur University who led the teams that collected the farm data; to Indrajit Mallick and Sandip Mitra

who helped us collect other relevant data. We are specially grateful to Asim Dasgupta for his encour-

agement, help and advice. For useful discussions and comments we thank Debu Bandyopadhyay, Abhijit

Banerjee, Eli Berman, Sam Bowles, James Boyce, Maitreesh Ghatak, Bhaswar Moitra, Sandip Mitra, T.N.

Srinivasan, MacArthur Inequality network members, and seminar participants at Brown, UC San Diego,

Monash, Stanford, Tufts, and the World Bank.2Department of Economics, University of California, Berkeley3Department of Economics, Boston University

1

1 Introduction

Land reforms represent policies with significant promise to increase farm productivity as well

as reduce rural poverty in developing countries. These reforms include land redistribution,

titling programs and regulation of tenancy contracts. A large theoretical literature going

back to Adam Smith, Alfred Marshall and John Stuart Mill has focused on the potentially

productivity enhancing impact of such reforms, a theme echoed in more recent literature

as well.4 The potentially beneficial incentive effects of these reforms stem from reduction

in sharecropping distortions, agency costs of hired labor, and improved access to credit.

On theoretical grounds however, the productivity effects are ambiguous as there are a

number of possible opposing effects as well, such as loss of scale economies, investment

capacity or incentives of owners, and reduced ability of owners to control their tenants.5

In addition, land reforms could have general equilibrium implications for local prices of

key factor inputs, the distribution of land across size classes and ownership forms, and the

nature of local governance that affect infrastructure and farm extension programs. Such

general equilibrium or governance impacts are less well understood or emphasized in existing

literature.

Empirical evidence on the productivity effects of land reforms has been provided in

recent years by Besley and Burgess (2000) and Banerjee, Gertler and Ghatak (2002) in

the context of Indian agriculture. They employ official government data at relatively high

levels of aggregation. Besley-Burgess examine the effects of land reform legislations on

growth and poverty reduction in a panel of different Indian states. Banerjee et al study

the effect of implementation of a given tenancy reform (Operation Barga) within the state

of West Bengal on rice yields in a panel of different districts, and interpret the effects in

terms of reduction of sharecropping distortions. Operation Barga registered sharecropping

contracts, protecting sharecroppers from eviction, and legislated minimum shares accruing

4For a short history of the classical debates on this question, see Johnson (1950). For a survey of the

more recent theoretical literature on sharecropping see Singh (1989) and Bardhan and Udry (1999). Also

see Binswanger et al (1993) for a survey of the literature on land reform in developing countries.5See Banerjee, Gertler and Ghatak (2002) for analysis of some of these contrasting effects of tenancy

regulations.

2

to the tenant.6

Compared to Besley and Burgess, the Banerjee et al paper pertains to a substantially

lower level of aggregation (districts rather than states) and uses a more reliable measure

of sharecropping reform (the proportion of tenants registered, rather than the number of

legislative amendments). Yet, the level of aggregation is still too high for the results to be

interpretable as the effects on sharecropping distortions alone. A more direct test would

require use of data at the level of individual farms, distinguishing between effects on tenant

and owner-cultivated farms, and between partial and general equilibrium effects. That is

the purpose of this paper.

We use a disaggregated farm panel in West Bengal, which has implemented a range

of reforms in land property rights and local governance since the late 1970s. This state

experienced a spurt in the rate of growth of agricultural production starting in the 1980s,

associated with large increases in the adoption of high yielding varieties (HYV) rice. We are

also interested in understanding the main causes of this growth; in particular the role of the

land reforms, as distinguished from farm extension programs, changes in market prices, or

autonomous processes of diffusion of HYV rice. Our analysis is distinguished from previous

studies by its use of more reliable farm production data, the level of disaggregation, controls

for other farm extension programs administered by local governments, and for potential

endogeneity of program implementation.

Our principal results are the following. Tenancy reforms had a statistically significant,

positive impact on farm productivity. However, we cannot precisely distinguish differences

in impact between tenant and owner-cultivated farms. There were large spillovers to non-

tenant farms, indicating the existence of a significant general equilibrium impact of the

tenancy reform. Hence the productivity impact of the reforms cannot be understood in

terms of reduction of sharecropping distortions in tenant farms alone. We explore a number

of possible channels of productivity spillovers from tenant to owner-cultivated farms. The

evidence is not consistent with various channels such as effects on the cost of credit, seeds,

land markets, improved targeting of extension programs by local governments, or diffusion of

6Further details of the program are provided in Section 2 below.

3

high-yielding rice varieties. It is possible, however, that they operated via local groundwater

markets: increased demand for water by tenants may have stimulated increased investment

in tubewells and thereby lowered the equilibrium price of groundwater. We are exploring

this channel in our currently ongoing research.

A second set of results concern the relative contribution to farm productivity growth of

land titling programs, the tenancy regulation program, and various farm extension programs

administered by local governments. We find no statistically significant effects of the land

titling program. The contribution of the tenancy regulation program is small, of the order

of 5%, in contrast to the estimate of 20% provided by Banerjee et al (2002). At the same

time the role of farm extension programs was much larger, ranging between 75% growth

over 1982–95 resulting from irrigation, and over 100% growth for supply of seeds, fertilizers

and credit.7 In particular the supply of agricultural minikits (containing HYV seeds and

fertilizers) had large effects on adoption of high-yielding rice varieties, whereas the tenancy

program had no significant effect on these adoption rates.8

Section 2 describes the context of West Bengal agriculture, the various land reform and

extension programs, and the nature of our data. Section 3 provides the main empirical

estimates of the programs on farm productivity. Section 4 explores possible channels of

impact, including direct tests for sharecropping distortion in tenant farms, and various

channels for spillovers to non-tenant farms. Section 5 concludes.

7These pertain to the effect of growth in these factors alone, which over-predict the actual growth ob-

served. There are many other controls in our regression, including common year effects and price changes,

which presumably caused actual productivity growth to be lower than is predicted by these input supply

programs alone.8Specifically, the bulk of the productivity impact of the tenancy reform consisted of increased yields of

traditional rice varieties, and similar increases in areas cropped under both traditional and new rice varieties.

4

2 Background and Data Description

2.1 Land Reform Programs

Following Independence in 1947, land reforms were an important priority for newly elected

governments at both the central and state levels in India. These included abolition of

intermediary landlords (zamindars), redistribution of lands above mandated ceilings, and

regulation of tenancy. Responsibility for agricultural policy is vested in state governments

under the Indian Constitution. Respective states proceeded to enact suitable legislation in

the early 1950s, with encouragement and assistance from the central government.

Like many other states, West Bengal faced a large number of problems with regard

to implementation of land reform laws until the late 1960s, owing to loopholes in legisla-

tion, poor land records and lack of political will (associated with the control of the state

government by the Congress party, whose rural base rested on the support of landlords).

From 1966 until 1971 a number of coalition governments were formed, in some of which

Left parties (including the Communist Party of India (Marxist), or CPM) played part. The

late 1960s also witnessed an armed peasant uprising in the state associated with extreme

Left-wing parties utilizing guerrilla tactics, which seriously threatened law and order in the

state. Partly for these reasons, elected governments in the 1970s were more resolved to

implement land reforms, by closing legal loopholes, and reinforcing administrative efforts.

A new Land Reforms Act was passed in 1971, redefining land ceilings that formed the basis

of land redistribution programs. In 1977 a Left-front coalition headed by the CPM won

an absolute majority in the state government. From the very beginning implementation of

land reforms was a top priority for this government, along with delegation of delivery of

development and welfare programs to a newly created three-tier system of directly elected

local governments. This Left coalition has subsequently remained in power at the state level

until the present day, thus enabling these reforms in land reform implementation and local

governance to be continued over successive decades.

There were two principal land reform programs. The first represents appropriation of

lands (a process known as vesting) above the legislated ceilings from large landowners, and

5

subsequent distribution of this land to the landless in the form of titles to small land plots

(called pattas). Most of the vesting had been carried out prior to 1978.9 According to

the Left Front government’s own admission, it had been unable to markedly increase the

extent of land vested over the amount available in 1978; hence its main initiative has been

the distribution of vested land in the form of land titles. Hence we shall focus on the

distribution of land titles to the poor. For the state as a whole, P.S. Appu (1996, Appendix

IV.3) estimates the extent of land distributed until 1992 at 6.72% of its operated area,

against a national average of 1.34%.10

The other land reform program involved registration of tenancy contracts. In 1977 the

newly elected Left Front government amended the 1971 Land Reforms Act, making share-

cropping hereditary, rendered eviction by landlords a punishable offense, and shifted the

onus of proof concerning identity of the actual tiller on the landlord. Subsequently they

initiated a mass mobilization drive with the assistance of farmer unions (Kisan Sabha) and

newly empowered local governments to identify sharecropping tenants and induce them to

register their contracts with the local Land Records office. Registration was also accom-

panied by a floor on the share accruing to tenants, amounting to 75% (replaced by 50% if

the landlord pays cost of all non-labor inputs). Over a million tenants were registered by

1981, up from 242,000 in 1978 (Lieten (1992, Table 5.1)), increasing to almost one and a

half million by 1990. Lieten (1992, p. 161) estimates on the basis of different assumptions

concerning the actual number of sharecroppers in the state, that upwards of 80% of all

sharecroppers were registered in the state by the early 1990s, while Banerjee et al (2002,

p.242) estimate this proportion to be around 65%.

These land reforms were complemented with creation of a three tier system of panchay-

ats or elected local governments (Gram Panchayats (GPs)), who were delegated responsi-

bility for delivery of various input supply services and local infrastructure. The principal

responsibilities entrusted to the panchayats included implementation of land reforms, of

9We were able to get data on the time pattern of vesting in 34 of our sample villages, where we found

70% had been vested prior to 1978.10Only one other state (Jammu and Kashmir) achieved a higher percentage, with the vast majority of

states distributing less than 1.5% of operated area.

6

the two principal poverty alleviation schemes (the Integrated Rural Development Program

which gave subsidized credit to the poor, and employment programs such as Food for Work

(FFW), National Rural Employment Program (NREP), Rural Labour Employment Guar-

antee Program (RLEGP) in the 1980s which were merged into the Jawahar Rozgar Yojana

(JRY) from 1989 onwards), distribution of subsidized agricultural inputs (in the form of

minikits containing seeds, fertilizers and pesticides), local infrastructure projects (including

roads and irrigation), and miscellaneous welfare schemes (old-age assistance, disaster relief,

housing programs for the poor etc.). The bulk of the funds (78% in our sample) for these

programs were devolved to the local governments under various schemes sponsored by the

central and state government. The funds percolated down from the central government to

GPs through the state government, its district-wide allocations, and then down through the

upper tiers of the panchayats at the block and district levels. Upper tiers of the panchayats

thus affected allocation across different GPs, while the main role of the GP was to select

beneficiaries of these schemes within their jurisdiction.

2.2 Data

Our study is based on data from cost of cultivation surveys carried out by the Department

of Agriculture of the state government. These surveys were carried out for the purpose of

estimating agricultural costs of principal crops in the state. These are aggregated at the

state level and eventually sent to the Commission for Agricultural Costs and Prices at the

central government in New Delhi, which uses this information to set foodgrain prices on a

cost-plus basis.

A number of reasons make the data from these surveys especially reliable. First, the

surveys are not used by the government to estimate agricultural production levels in the

state. So they are not subject to reporting biases that have been argued to afflict published

statistics of the state government used by most previous studies (including Besley-Burgess

(2000) and Banerjee et al (2002)).11 Second, the surveys selected a stratified random sample

11Considerable doubt has been raised about the reliability of agricultural output data of the West Bengal

state government, by Boyce (1987) and Datta Ray (1994). They describe how the West Bengal state

7

of farms in West Bengal, first randomly selecting blocks within each district, then selecting

pairs of neighboring villages randomly within blocks, and finally selecting a random sample

of 8 farms in each village, stratifying by landholdings.12 Every five years the samples were

redrawn and freshly chosen. Third, each selected farm was visited on a bi-weekly basis for

five successive years. Trained investigators measured principal outputs and inputs of farms

on a weekly basis, and every year filed an assessment of costs on various items, following

prescribed norms by the agriculture department. Prices of main inputs and outputs were

also collected at the farm-year level.

Detailed farm records could be located for three successive five year panels, spanning

1981–96. Within each five year panel, complete data is available for all farms in each village

for an average of four years. We thus create an unbalanced panel covering 89 villages

located in all but two districts (Kolkata and Darjeeling, neither of which has any significant

agricultural activity).13

The farm data is complemented by village data collected from a variety of sources. We

carried out household surveys in these villages, and collected data relating to composition

and activities of local governments, spanning the period 1978-2004. These enable us to

assess changes in the distribution of land, literacy, caste and occupations in the village;

composition of elected GPs; details of infrastructure programs and yearly budgets of these

GPs. We visited local land reform offices to obtain data on yearly land reform imple-

mentation (land titles distributed and tenants registered, including names of beneficiaries

and cultivation areas involved). Visits to local lead banks and block development offices

government has often shifted between agricultural statistics collected from sample surveys and crop cutting

surveys initiated by Mahalanobis in the 1940s, and those based on subjective ’eye estimates’ from the state

Directorate of Agriculture. These concerns are aggravated by the frequent use of published statistics by

the West Bengal state government to claim credit for their policies in generating a high rate of agricultural

growth during the 1980s and the first half of the 1990s.12However we use a sub-sample of the original sample, based on the village-years for which complete farm

records could be located in local offices of the state agriculture department. We do not know whether this

may have caused any bias.13Darjeeling occupies a hilly terrain at the foothills of the Himalayas. It has a number of tea estates and

plantations, but grows no principal foodgrains, cash crops (apart from tea) or vegetables.

8

generated yearly data on distribution of IRDP credit and agricultural minikits in each vil-

lage. GP records yielded yearly allocation of spending and scale of various infrastructure

projects. We also collected data on rainfall from local recording centers of the state Me-

teorological department, leading economic indicators at the district or regional level from

published statistics of the state government, and outcomes of elections to the state and

national legislatures in each constituency spanning the sample areas.

Summary statistics concerning the villages in our sample are provided in Table 1. The

sample includes 89 villages in 57 GPs. Each GP consists of ten to twenty elected members

of a council governing administration of the jurisdiction of the GP, which usually consists

of eight to fifteen villages or mouzas. On average each district comprises 20 blocks and

200 GPs. Table 1 shows the principal demographic and asset distribution changes in the

sample villages between 1978 and 1998, based on an indirect household survey administered

in each village in 1998.14 The distribution of cultivable non-patta land (i.e., excluding land

distributed through the land reforms) shows increasing landlessness, as well as a redistri-

bution of land from large to small landowning households, resulting from splitting of large

landholdings (via market sales of land and household sub-division).

Table 2 indicates the extent of land reform implemented in our sample by 1998. Ap-

proximately 5.4% of cultivable land was distributed to 15% of the population in the form of

registered land titles. Approximately 6% of cultivable land involved leased lands on which

tenants (bargadars) were recorded. However, 2% had already been registered by 1978, so

the incremental area covered by the program since 1978 was 4%. The proportion of house-

holds registered by 1998 was 4.4%. Aggregating the two programs, therefore, about 11.5%

of operational land area was affected, and 20% of all households benefited.

Table 3 depicts trends in agricultural inputs provided by the GPs in our sample vil-

14This involved selection of voter lists for 1978 and 1998, and interviews with four or five different senior

citizens in each village to identify land and demographic status of each household for those two years

respectively. The details of this survey are described further in Bardhan and Mookherjee (BM, hereafter)

(2004b). In particular the land distribution obtained thereby when aggregated to the district level matches

quite closely the distributions reported in the state Agricultural Census as well as National Sample Survey

(NSS) decadal surveys of operational holdings in West Bengal.

9

lages between 1982–95.15 The 1980s witnessed larger supplies of IRDP credit and minikits

compared with the 1990s. One out of every nine households received minikits in the 1980s,

containing seeds, fertilizers and pesticides. The bulk of employment funds were spent by

GPs on building and maintenance of local roads; these employment programs created 2-4

mandays of employment per household every year. There was also expansion of areas irri-

gated by state canals. Greater expansions were witnessed in medium and small irrigation

projects, many of which were managed by panchayat officials. Spending on irrigation by

the panchayats was highest during the early part of the period, fell sharply throughout the

1980s, and stabilized thereafter. A similar trend was observed for GP spending on local

roads, except that there was an upturn towards the end of the period.

The last two rows of Table 3 provide (for the sake of comparison) the land reforms imple-

mented at these corresponding years: these peaked in the first half of the 1980s and tailed

away thereafter. Clearly, these became insignificant relative to input and infrastructure

provision from the mid-1980s onwards.

Table 4 shows average allocation of cropped area across different crops in our farm

sample, and their respective yields (measured by value added per acre). Rice accounted

for two-thirds of cropped area, with HYV rice accounting for 28% on average across the

entire period. HYV rice yields were two and a half times those of traditional rice varieties.

Only potatoes generate a higher return (measured by value added per acre) than HYV rice;

however the short potato season (which lasts 70-90 days) limits the acreage devoted to this

crop. Other cash crops such as jute and tobacco generate high returns, followed by pulses,

vegetables and oilseeds, with wheat generating the lowest returns.

Table 5 shows changes over time in cropping patterns, yields and incomes: these are

weighted averages across farms in the sample for the beginning and end of each separate

panel (1982–85, 1986–90 and 1991–96).16 Cropped area per farm grew at a modest rate

15These averages use operational land area in different villages as weights, the reason the numbers reported

here differ from the unweighted averages reported in BM (2004a, 2006).16Cropped area per farm and hired labor employed are not weighted. The three panels contained 56, 155

and 131 farms respectively. We do not report averages for the first year 1981 of the first panel, owing to

many missing observations for that year.

10

(about 10%) for each five year period in the 1980s, but did not grow in the 1990s. HYV

rice adoption rates grew dramatically, from 1 to 8% in the first panel, 16 to 32% in the

second panel, and 46 to 53% in the third. Value added per acre in rice grew analogously,

by almost 50%, 80% and 30% in the three panels. Farm incomes grew at similar rates,

indicating the importance of rice in the growth process. The wage rate of hired workers

remained stationary throughout the 1980s, but grew about 20% in the first half of the 1990s.

Employment rates increased during the 1980s, but fell by about 15% during the early 1990s.

Hence incomes of agricultural workers, the poorest section of the rural population, grew at

a modest rate.

Table 6 shows the fraction of farms which leased in land, and associated areas leased in,

for different years in the sample, averaged across all the farms in the sample for any given

year. This is based on the cost of cultivation survey: we infer if a farm is leasing in land for

production of a specific crop in any given year if some rent is paid to a landlord.17 In the

first panel lasting until 1985, there seems to be a downward trend in the extent of tenancy,

with the percent land area leased falling from 13% to 7%. In later panels (1986–90, 1991–

95) no trends are visible, averaging between 1–2% during 1986–90 and around 6% during

1991–95. The proportion of farms leasing in land seems rather low, ranging between 1–2%.

In comparison Table 2 calculated on the basis of the village land records and the indirect

household survey shows that 3–4% of households in the sample villages were registered in

the program. So there is possibly some under-representation of tenants in the sample.

Further information concerning the nature of tenancy contracts is not available in the

cost of cultivation surveys. For the state of West Bengal as a whole, the Operational

Holdings survey of the National Sample Survey (NSS) for the year 1991-92 indicates that

14% of all operational holdings (and 10.4% of the area) was leased in. Of these 3.7% were

fixed rent tenants, while 8.8% were sharecroppers. Of the total area leased in, about 48%

was on sharecropped contracts, and 19% on fixed rent contracts. Hence on the basis of the

NSS estimates the extent of sharecropping tenancy in the state seems rather low, of the

17Indeed, we do not know what part of the land devoted to that crop was leased in. So we assume that

all of it was. This implies that the tenancy area estimate is biased upward.

11

order of 5% of operational area. It seems unlikely on this basis that the tenancy registration

program could explain a rise in farm yields of the order of 20% estimated by Banerjee et al

(2002), or a rise in the adoption rate of HYV rice to over 30% of total cropped area by the

mid-90s.

The NSS 1991-92 survey also indicates two thirds of all tenancy leases exceeded two

years in duration, and 48% exceeded five years in duration. Short term leases of a year

or less accounted for 22%. Details concerning tenant crop shares are not available from

any source, apart from the survey carried out by Banerjee et al (2002) for a sample of 20

villages, who report that the proportion of tenants with crop shares exceeding 75% rose

from approximately 5% to approximately 20% with the reforms, and those exceeding 50%

rose from 17% to 39%. There is a significant peak at a crop share of exactly 50%, which

fell from approximately 80% to approximately 60%. Hence the program appears to have

raised crop shares for approximately 20% of the tenants. This reduces the incidence of the

program even further. In summary, the extent of area covered by the tenancy registration

program appears quite limited, and there is some under-representation of the incidence of

tenancy in our sample compared to the entire state.

3 Estimating Effects of Programs on Farm Yields

3.1 Effect of Land Reforms: Theoretical Hypotheses

The effect of land reforms on farm productivity have been the topic of a large literature

in development economics. The classic arguments concern Marshallian inefficiencies arising

from sharecropping, where the share paid to the landlord acts as a tax on the tenant’s ef-

fort. Sharecropper registration can raise farmer incentives by capping this implicit tax rate.

Other incentive effects arise from removing the right of landlords to evict tenants: the direc-

tion of these are ambiguous, owing to conflicts between different effects.18 Eviction threats

can be used by landlords as an incentive device, the removal of which could dull tenant

18Regulation of eviction rights may be needed in the absence of the ability of landlords to commit to leases

of long duration.

12

incentives. On the other hand, security of tenure may promote longer time horizons for the

tenant and thereby increase investment incentives. These issues are discussed in Bardhan

(1984), Dutta, Ray and Sengupta (1989) and Banerjee, Gertler and Ghatak (2002).19 In

addition, registered sharecroppers were eligible to apply for production loans from formal

credit channels, which could reduce their credit costs (owing to significant differences in

interest rates between formal and informal credit sources).20

The incentive effects of redistributing land ownership have also been discussed in pre-

vious literature (Bardhan (1973), Berry and Cline (1979), Eswaran and Kotwal (1986),

Binswanger et al (1993), Mookherjee (1997)). In general, the effect depends on the extent

of economies or diseconomies of scale. Given the advantages of family labor cultivation over

hired labor, and the relative lack of important sources of scale economies (such as mecha-

nization) in rice cultivation, one might expect small farms to be more productive than large

farms. While such a pattern has frequently been empirically observed, it is possible that

they reflect differences in unobserved soil characteristics between small and large farms. If

more productive lands are more prone to fragmentation, small farms may be expected to

have more fertile soils, in which case observed yield differences between small and large

farms overstate the effect of land redistribution programs.

As explained previously the land titling program mainly concerned distribution of ti-

tles to land that had previously been appropriated from those holding surplus land above

legislated land ceilings. Distribution of already-vested lands would enable them to be ac-

tively cultivated instead of lying fallow, in which case one would expect a rise in production

yields. Of course these yield improvements would be negligible if the transferred lands

were of inferior quality or of very small size. The average size of land parcels distributed

19The theoretical literature (see, e.g., Mookherjee (1997)) also clarifies that the reforms can raise farm

productivity but typically do not constitute a Pareto improvement. Landlords and those with more land

than the ceiling allows are typically rendered worse off, while tenants and recipients of land titles are better

off. This also explains why sharecropping as an institution persists, despite the existence of associated

distortions. Voluntary sales of land to tenants or landless households would therefore not occur in a laissez

faire economy.20We have learnt this from interviews with government and bank officials, as well as sharecroppers. We

do not, however, have data on access and costs of credit.

13

was approximately half an acre, compared with an average size of 1.5 acres for plots regis-

tered under Operation Barga, the tenancy registration program. Moreover, while the latter

were cultivable by their very nature, approximately half of all titles distributed consisted

of non-cultivable land. We have also been told by bank officials and farmers that we inter-

viewed that farmers were not eligible for bank loans on the basis of the titles received in the

land reform program, mainly owing to the uneconomically small size and poor quality of

the land parcels concerned. Therefore the productivity impact of the land titling program

could be expected to be less significant than the effects of tenancy registration, and we focus

principally on the latter.

3.2 Impact of Operation Barga on Farm Value Added Per Acre

Column 1 in Table 7 presents an OLS panel regression of farm value added per acre on

a dummy for farms leasing in land, and on the percent of land registered in the village

until the previous year, apart from farm, year dummies and controls for farm acreage. 21

Note that the tenancy dummy of obtained from the farm cost of cultivation data, while the

tenancy registration rate is obtained from the land records office pertaining to the village

as a whole. The reason we do the latter is that we do not know if the tenant farmer in

question had registered in any given year. Hence we use the registration rate in the village as

a proxy for the likelihood that a randomly chosen tenant farmer has been registered. Even

if a particular tenant did not register, the option to register would be expected to raise the

bargaining power and hence the share accruing to the tenant. Hence the registration rate

in the village is also a reasonable indicator of this outside option effect.

To control for possible endogeneity of the leasing decision, the second column replaces

the current lease dummy by the fraction of years in the sample (that the farmer appears)

that it leased in land. By construction this is a farm-related variable that does not change

21We know from some of the literature on agency problems with respect to hired labor in conjunction

with credit market imperfections may cause farms to rely on family labor as far as possible (Eswaran and

Kotwal (1986)). Given family size, increases in cropped area cause increasing reliance on hired labor, which

therefore tends to increase agency problems and lowers farm profits. We therefore need to include controls

for cropped area in the regression to capture this effect.

14

across years, and represents differences in the average extent of leasing across farms. The

interaction of this variable with the extent of tenancy registration in the village represents

variations in the effect of the program across tenant and non-tenant farms.

These two columns in Table 7 show that the effect of leasing and its interaction with

tenancy registration rates in the village have the signs that would be expected from the

Marshall-Mill theory of sharecropping distortions, but these effects are statistically insignif-

icant.

The next set of columns in Table 7 include possible spillover effects of the tenancy

registration program on non-tenant farms, as well as controls for the land titling program

and other extension programs implemented by local governments. We add the tenancy

registration rate as a regressor, besides its interaction with leasing. Note that we use the

percent area registered as estimated from the local land records, rather than the proportion

of tenants registered (the measure used by Banerjee et al). The proportion of area registered

obviously represents more closely to the incidence of the program in the village, compared

with the proportion of tenants registered. The latter proportion averages around 50% in

the sample, and 65% for all West Bengal in 1993, in contrast to less that 5% operational

area that was covered by the program.

Note also that the registration rate pertains to the village as a whole: this regressor

by itself (i.e., controlling for its interaction with the lease dummy or fraction) represents a

common effect of the program on the profitability of all farms in the village in future years,

via general equilibrium, governance or learning spillovers.

The regressions in columns 3,4 of Table 7 also add as controls the distribution of land

titles to poor households, and various farm input supply and infrastructure programs ad-

ministered by GPs: specifically, cumulative lagged values of: (a) proportion of cultivable

land distributed in the form of pattas; (b) minikits distributed in the village per household,

(c) IRDP credit subsidy delivered per household, (c) log of the cumulative GP expendi-

tures (in constant 1980 prices) per household on local irrigation and road projects, and (d)

cumulative mandays of employment generated by GP programs per household. Additional

controls include annual rainfall at the nearest weather station, the log of the rice price

15

received by the farmer, canals and roads provided by the state government in the district,

apart from farm and year effects. The regression specification is thus:

Vfvt = β1Lfvt+β2Lfvt∗Bv,t−1+β3Afvt+β4A2fvt+β5Bv,t−1+β6Pv,t−1+β7Ev,t−1+γf +δt+εfvt

where Lfvt denotes the dummy for whether farm f in village v in year t leased in land,

Bv,t−1 denotes cumulative proportion of agricultural land registered under Operation Barga

in village until year t − 1, Afvt is the gross cropped area in the farm in year t, Pv,t−1 is the

cumulative proportion of cultivable land in the village distributed in the form of land titles,

Ev,t−1 is the cumulative per capita delivery of extension services to the village until year

t − 1. The coefficients β5, β6, β7 represent spillover effects of village development programs

delivered by GPs on future profits of all farms in the village, in contrast to β2 which is the

direct effect on farms leasing in land owing to a reduction in sharecropping distortions.

We see now that inclusion of the village controls representing different government in-

terventions results in a higher coefficient on the interaction of leasing with Operation Barga

implementation.22 However, in no version is this coefficient statistically significant, even at

10% significance level.

More surprising is the strong significance of the spillover effects of the tenancy regis-

tration program, the delivery of kits, credit and local irrigation. These represent common

effects of these interventions on profitability of all farms in the village, most of which we

know (from Table 6) were not leasing any land at all. The estimated spillover effect of

tenancy registration on non-tenant farms is larger than the point estimate of the direct

impact on the tenant farms.

22This reflects a negative partial correlation between Barga implementation and delivery of minikits, the

program which we shall see was the largest significant source of farm productivity growth. A regression of

% land registered on all the other programs, controlling for village and year effects, shows a statistically

significant (at 1%) and negative effect of cumulative minikits delivered per household. In other words,

Operation Barga tended to be implemented more vigorously in villages with relatively slow growth in minikit

delivery. Hence controlling for the latter results in an increase in the estimated effect of Operation Barga.

16

3.3 Controls for Endogeneity of Program Implementation

We next examine robustness of the estimates in Table 7 to possible endogeneity of program

implementation. Banerjee et al (2002) argue that implementation of tenancy registration

was largely exogenous, driven by myriad bureaucratic factors external to the villages con-

cerned. Nevertheless to the extent that registration was ultimately demand-driven (being

based on a voluntary decision of tenants to register their contracts, and efforts of local gov-

ernments to implement the program), there is always a danger of possible reverse causality,

both at the farm and village level. Specifically, farms or villages that seek to attain higher

rates of productivity increase may be more inclined to register themselves or their tenants.

Our previous work on the political economy of the land reforms (BM (2004a,b)) indi-

cates the extent of political competition between the Left Front (LF) and its principal rival

during the study period, the Indian National Congress (INC), played a role in explaining

temporal village-specific variations in implementation rates. In particular, increased com-

petition resulted in higher implementation rates: there was an inverted-U pattern between

implementation and the proportion of GP seats secured by the LF, in a village panel re-

gression of implementation rates. The proportion of seats won by the LF in any given GP

election was related, in turn, to a number of factors affecting relative standing of the two

parties with voters in the concerned district. These include variables arguably exogenous

with respect to time-varying village-specific farm productivity: the presence of the INC in

the Lower House (Lok Sabha) of the National Parliament in New Delhi, the annual infla-

tion rate of the CPI in the region in question, and the rate of new manufacturing factory

employment in the district, in addition to (and interacted with) local incumbency (fraction

of seats won by the LF in the previous GP election).23

Accordingly we use as instruments for program implementation the above district, re-

gional and national level determinants of relative competitive strengths of the LF and INC,

their squares, and interactions of these with lagged LF share in the GP. The underlying

identification assumption is that year-to-year fluctuations in these were uncorrelated with

23An Arellano-Bond specification for dynamics of LF share of GP seats is not rejected, implying lagged

shares are a valid instrument for current shares after controlling for village fixed effects.

17

farm productivity in each village, after accounting for their effect on various programs

affecting productivity of local farms. The validity of this assumption depends on the com-

prehensiveness of the set of programs that we include as controls.

One cannot be completely sure of course that all relevant direct channels of influence

have been controlled for. But our IV estimates should remove a large source of local

time-varying sources of unobserved village-level factors affecting farm productivity growth,

predicting them instead on the basis of political and economic factors on a substantially

wider geographic area (interacted with local incumbency patterns). The direction of bias

associated with potential endogeneity of implementation rates at the local level can be

inferred by comparing OLS and IV estimates.

Other extension programs are instrumented by the scale of the corresponding program

at the level of the state, interacted with the same variables affecting political competition.

The idea here is that temporal fluctuations in the scale of these programs in the state as a

whole resulted from financial, administrative and political factors at the level of the state

government. These percolated down to villages and farms within those villages through a

hierarchical system, with relative allocations across different villages determined partly by

the same variables representing political competition at the district or regional levels.

We use this method to control for endogeneity of the three principal programs: IRDP

credit, minikits, and sponsored program grants devolved to GPs out of which local irrigation

projects are funded. The scale of the program at the state level is estimated by the average

supply of the program across all sample villages in any given year.

The detailed set of instruments is therefore the following: lagged cumulatives of (a)

percent Congress seats in Parliament in the previous General Election, interacted with

lagged Left share of GP seats; (b) local inflation rate of cost of living, interacted with lagged

Left share of GP seats; (c) lagged Left share of GP seats; (d) squares of the above variables;

(e) interactions of the previous variables with IRDP credit, minikits, and sponsored program

funds, averaged across all GPs in that year. F-statistics and partial-R sq. for first stage

regressions are reported at the bottom of the Table, all of which are significant.

Column 1 of Table 8 shows the IV estimates of the regression previously reported in

18

column 3 of Table 7. The IV estimate of the spillover effect of the tenancy registration pro-

gram, and of the minikits distributed, is approximately twice the size of the OLS estimate.

Both are statistically significant at the 1% level. The IRDP credit disbursement effect also

gains in significance, and now the employment mandays generated by local infrastructure

programs also becomes significant at the 5% level.

The fact that the IV estimates are substantially larger than the OLS estimates possibly

reflects the tendency for land reform and farm extension programs to be more vigorously

implemented in villages where other sources of farm productivity growth were lower, i.e.,

they were implemented more vigorously in lagging regions. This may have owed to the

efforts of the concerned local governments to speed up development of the lagging regions,

and the greater desire of farms in these regions to catch up with those in other regions.

Even if one may doubt whether the exclusion restrictions are exactly valid for the set of

instruments we use, the fact that the OLS results are strengthened in the IV regressions

indicates that it is unlikely that the were driven by endogeneity of program implementation.

The estimated differential impact of tenancy registration on yields of tenant farms vis-

a-vis owner-cultivated farms is however unchanged; it continues to remain insignificant

statistically. Hence controlling for endogeneity of program implementation appears to raise

the spillover effect of tenancy registration, without enabling us to precisely measure the

direct incentive effects of the reforms. To doubly verify our interpretation of the spillover

effects, columns 2 and 3 of Table 8 re-run the same regression in OLS and IV for the sample

consisting of pure owner-cultivated farms, and find the same results.

Next, Table 9 reports corresponding OLS and IV regressions at a higher level of aggrega-

tion: average yields in the village (computed using the village sample of farms and weighting

their respective yields by their relative cropped areas). This allows us to study heterogene-

ity of treatment impacts on (and changes in composition within the village across) farms of

differing sizes. It controls for one set of possible general equilibrium impacts: changes in the

distribution of cropped area in the village between marginal farms (with operational holding

below 2.5 acres), medium (between 2.5 and 5 acres) and big farms, and the proportion of

these respective areas operated by tenant farms. Interactions of the latter with the extent

19

of tenancy registration in the village are reported in the first three rows, thus allowing the

direct incentive effect of the reform to depend on farm size.

We see that the OLS estimate of the direct effect of the reform is significant (at the 10%

level) for medium sized farms, and is substantially larger than the previous estimates of the

average (direct) effect on all tenant farms. For medium sized farms the direct effect is also

substantially larger than the spillover effect. This suggests substantial heterogeneity across

different size categories in the direct productivity effects of Operation Barga on tenant

farms. However, the IV estimate of the direct effect is statistically insignificant, though

of similar magnitude as the OLS estimate. Hence the direct effects of the tenancy reform

continue to be imprecisely estimated.

At the same time, the spillover effects continue to be large and significant, though

somewhat smaller in magnitude than the farm-level estimate. Presumably induced effects

of the programs on the distribution across farms within the village cause some reduction in

the productivity effect, but they are not large enough to nullify the increase in intra-farm

productivity.

To assess the relative quantitative significance of the direct and spillover effects, as well

as of the land reforms relative to farm input services, Table 10A calculates the predicted

impact of different programs on farm yields. We calculate the percent change in value added

per acre between 1982 and 1995 in a hypothetical village in which the proportion of land

registered under Operation Barga was equal to the weighted average for the entire sample

(with weights proportional to operational areas of cultivation). We compare this with the

change predicted by a hypothetical change in cumulative supply of kits, credit and GP

expenditures on irrigation, equal to the weighted average of the observed changes in the

sample villages (with weights again taken proportional to operational areas). We calculate

the impact of tenancy registration to be of the order of 5% for yields on non-tenant farms.

This estimate would be higher for tenant farms, depending on which estimate of the direct

productivity impact on tenant farms we take. Tables 7 and 8 indicates the direct impact

is smaller than the general spillover effect, while Table 9 indicates it may be two to three

times larger for medium-sized farms. This indicates a range of 8 to 20% increase in yields

20

for tenant farms.

The predicted impact of Operation Barga are dwarfed by those of minikits (over 500%),

IRDP credit (over 100%), and matched by local irrigation (6%). The similarity of corre-

sponding regression coefficients suggests that the social rate of return to tenancy registra-

tion was comparable to those of other input services. The differences in measured impacts

principally reflect differences in the scales of these respective programs.

Table 10B presents more detailed estimates of the predicted impacts of different pro-

grams implemented, using actual changes in program implementation in each village for

the duration of each farm sample separately. For each five year time block in which any

given village appears in the sample, the change predicted in farm yields in those villages by

the actual changes observed in different programs is first calculated, using operational land

areas of different farms (averaged across different years) to weight different farms. This is

subsequently averaged across different villages in the same time block, using their relative

operational areas as weights. The broad results remain unchanged from Table 10A. The

impact of the tenancy program remains small, that of credit and kits remains roughly the

same, and that of GP irrigation expenditures becomes larger. Most of the strikingly large

impacts are predicted for the first five years 1981–85, with subsequent impacts tailing off

subsequently.

4 Understanding Channels of Impact

In this section we try to understand possible channels of impact of the various programs. To

assess the nature of the direct incentive effects of the tenancy reform, we first review evidence

concerning the significance of sharecropping distortions by comparing input applications

between leased and non-leased plots. Thereafter we seek evidence concerning the nature of

the spillover effects of the tenancy reform on non-tenant farms. We explore the plausibility

of various channels, ranging from prices of credit, seeds and water; under-reporting of

tenancy; induced changes in land distribution or rental markets; improved targeting of

extension programs by local governments; and social learning in the process of diffusion of

21

HYV rice varieties.

4.1 Sharecropping Distortions

The principal finding above is that the direct incentive effects of the tenancy reform were

weaker and less precisely measured, compared with the spillover effects. One can seek

evidence in alternate ways of the importance of sharecropping distortions, by comparing

inputs applied in tenant and non-tenant farms, using the methodology of Shaban (1987)

and Braido (2006) to compare factor allocations across sharecropped plots and owned plots

in ICRISAT data from South and Central India. Table 11 compares the ratio of input

expenditures to the value of output between leased and non-leased plots in HYV and non-

HYV rice cultivation respectively. The regressions include farmer and year dummies: hence

they compare inputs applied by the same farmer between leased and non-leased plots in

the cultivation of the same crop. Since the dependent variable divides input expenditures

by the value of output, this controls for heterogeneity between leased and non-leased plots

(assuming a Cobb-Douglas specification of the production function, as explained in further

detail by Braido (2006)).

The first column of Table 11 shows that farmers significantly under-applied (by about

20%) inputs to leased plots compared with owned plots in HYV rice cultivation. However

the third column shows no significant difference in non-HYV rice cultivation. The second

and fourth columns interact the leased dummy with the extent of land registered in the

village. In the case of HYV rice, the extent of under-application fell by 12% (as a result

of 100% registration), but this is statistically insignificant. Since the area registered was

substantially smaller than 100%, the evidence shows little direct impact on incentives of

farmers to apply inputs on leased plots. In the case of non-HYV rice, tenancy registration

has a negative, insignificant effect on the application of inputs in leased plots. Hence we

have some direct evidence of the existence of sharecropping distortions but this is limited to

HYV rice, and even there we obtain very limited evidence that it was reduced significantly

by the tenancy registration program.

22

4.2 GE Effects: Credit or Seed Prices

One possible channel of general equilibrium impact on the village economy might operate

via induced changes in the demand and hence prices of key factors such as credit, seeds

or fertilizers. For instance, if registration of tenancy status rendered sharecroppers eligible

for credit from banks at interest rates substantially below informal interest rates, informal

interest rates within the village for crop loans may have declined. This could generate a

spillover to owner-cultivated farms in the same village. Unfortunately the farm data does

not include information about actual interest rates paid by farmers, so we cannot directly

check for pecuniary externalities through the informal interest rate.

One would expect smaller, poorer farmers to rely more on informal credit, in which case

the increases in productivity ought to have been larger for small farms. Table 12 examines

differential effects on productivity and incomes of small (less than 5 acre) and marginal (less

than 2.5 acre) farms. These are insignificant, both quantitatively and statistically. Therefore

there is no evidence of a larger impact on smaller farms. This suggests that impacts on

credit market imperfections are unlikely to constitute the source of the measured spillovers.

An alternative indicator of lowered costs of informal credit are the prices of seeds paid

by farmers, since informal credit is frequently bundled with purchase of seeds at inflated

prices (either directly from lender-traders, or from traders in a triadic relationship with

lenders and farmers). The farm data includes reliable data on seed prices actually paid for

HYV and non-HYV rice for a reasonably large sub-sample (i.e., containing more than 50%)

of farms growing these crops. In regressions of rice seed prices on programs implemented

(not reported here), we found no evidence any of the three programs on seed prices.

4.3 Effects on Local Governance and Targeting of Farm Services

The tenancy reforms may have reduced the extent of capture of local government by landed

elites, thus serving to target farm services better to small and marginal farms.24 If smaller

24This hypothesis has been advanced in the literature on fiscal decentralization (see, e.g., Bardhan and

Mookherjee (2000)), and is frequently invoked by representatives of the Left Front government in West

Bengal. However, we have no direct evidence concerning the effects of land reform on targeting, owing to

23

farms are more productive than large farms25, this may conceivably raise productivity in

the former group, as well as on the average farm.

However, there is no evidence consistent with this hypothesis. Table 12 showed that

the impact on small or marginal farms did not significantly exceed that on larger farms.

Moreover, in Tables 7 and 8 when we added controls for the proportion of various farm

services delivered to small and marginal farms, it turned out to have insignificant impact,

and did not alter any of the results reported in those tables.

4.4 Undisclosed Tenancy

The measured spillovers could conceivably reflect undisclosed tenancy: in the presence of

sharecropper regulations some landlords may lease land to tenants only if the latter did

not disclose the leasing relationship.26 We have noted in Section 2 that our sample seems

to under-represent the extent of tenancy, by comparing with the land records or NSS data

for the state as a whole. It is conceivable therefore that some farms claiming to be owner-

cultivated were actually tenant farms. The presence of the registration program is likely to

raise the bargaining power of these tenants, so some of the measured spillover effects may

reflect the positive incentive effects of this.

However, we remain skeptical of this possible explanation. Given the political climate

within which these land reforms were implemented, i.e., with LF-led local governments

hostile to the interests of large landowners, the tendency or such under-reporting should

be more pronounced for small farmers leasing in land from big landowners. We would thus

expect a larger measured impact on small and marginal farms. But Table 12 showed that

the spillover effects did not differ significantly across farms of varying size.

the difficulty of statistically identifying this effect in a convincing manner.25Indeed, we do find in our sample that farm value added per acre is declining significantly in farm cropped

area26We thank Maitreesh Ghatak for this observation.

24

4.5 Induced Changes in Land Distribution

An alternative general equilibrium impact of tenancy reforms could be on the distribution

of land, via land markets or household fragmentation. However, these would operate by

affecting the distribution of cultivated land across farms of varying size and ownership

type, and thus cannot explain changes in the productivity within farms that have been

documented in the previous section to form the bulk of the overall effect. Such effects could

arise only if the programs affected the extent to which individual farms tended to lease in

land, or by affecting the size of a representative farm (and of course, provided leasing or

size had significant productivity effects).

Table 6 indicated no pronounced trends in the extent of leasing. Exploring this issue

more carefully, we regressed the proportion of farmers leasing, or area leased within the

village, on the various programs implemented, after controlling for farm, year dummies and

other controls used in Tables 7 and 8. None of the programs displayed any significant effects.

The regressions in Tables 7 and 8 indicate a significant negative relationship between

cropped area and farm productivity, after controlling for farm and year dummies, and all

the other controls in those regressions.27 Is it conceivable that the tenancy program raised

farm productivity by inducing a shrinkage in cropped area of the representative farm? Note

first that this cannot explain the results of Tables 7,8 because those regressions included

farm size among the controls.

Moreover, there is evidence that the tenancy registration program induced an expansion

of cropped area. Table 13 presents IV regressions of gross cropped area in the farm as

a whole, as well as to HYV rice, non-HYV rice and potatoes, on the various programs

implemented, controlling for farm and year dummies. The tenancy program induced a

significant net expansion of cropped area, represented principally with increased cropping

of both traditional and HYV rice varieties. There is no significant impact of any of the

farm extension programs on aggregate cropped area. Hence there is no evidence that the

27Farm cropped area and its square formed part of the controls in those regressions, whose coefficients

were not reported there. A doubling of cropped area was associated with a 10% reduction of farm value

added per acre, which was statistically significant at the 1% level in all the regressions in Tables 7,8.

25

programs raised farm productivity on account of shrinking farm size or cropped area.

4.6 Adoption and Diffusion of HYV Rice

Table 13 also provides useful information about the impact of the programs on adoption of

HYV rice. The tenancy program was associated with a uniform expansion of areas allocated

to both traditional and HYV rice, at the expense of area allocated to potatoes. Hence the

evidence does not show a significant impact of Operation Barga on HYV rice adoption

vis-a-vis traditional varieties, after controlling for farm and year dummies. It is therefore

unlikely that social learning in the process of diffusion of HYV rice constituted an important

source of spillover.

On the other hand, the minikits distributed induced a significant expansion of HYV rice

area, as well as area allocated to potatoes, and an insignificant effect on non-HYV rice area.

Therefore the minikit distribution program seems to have increased incentives to adopt

HYV rice and high-value cash crops such as potatoes (refer Table 4) vis-a-vis traditional

rice varieties. The estimated impact of minikit supply on area allocated to HYV rice was

also ten times larger than the effect of tenancy registration.28 This probably accounts for

the larger productivity impact of the minikits than the tenancy reform.

4.7 Irrigation

Finally, tenancy reform may have affected access to groundwater irrigation. It is well

known that the period covered here displayed a substantial increase in irrigation. This was

partly on account of the high water needs of the summer boro rice crop, which is mostly a

HYV crop. Table 14 shows that farm area irrigated by shallow tubewells29 rose from 7.6%

28With 10% land registered, the area allocated to HYV rice increased by approximately .07 acres. The

increase in minikits supplied was of the order of 1 minikit per household, implying an increase in HYV rice

area by approximately 0.7 acres.29The shallower ones among these consist of wells with a depth of 30–50 feet, costing Rs. 11-12,000 each.

The deeper ones, referred to as submersible or min-deep tubewells have an average depth of 180 feet and

cost Rs. 50,000 each. These are operated with either diesel pumpsets or electricity.

26

in 1981 to over 31% in 1995, while from government canals rose from 5.2 to 5.8%, deep

tubewells30 rose from 2.8 to 4.5%, and ponds or river lift schemes (usually constructed by

local governments) rose from 7.5 to 14.8%.31

Shallow tubewells represent a form of (mostly) private investment in irrigation. The

investment required in a shallow tubewell is typically beyond the capacity of small and

marginal farmers, who purchase water from larger farmers that can afford to invest in

them. Detailed case studies in Moitra and Das (2004) of a number of villages indicate that

in the early 1980s the government provided subsidies and loans to farmers in order to invest

in shallow tubewells. Later in the 1980s as the water depth fell, investments switched to the

deeper submersible wells or dugwells which were more expensive, and in which most of the

investments were private and unsubsidized. The significant expansion in these tubewells

(combined with their role in expanding boro rice cultivation) have raised questions among

various commentators about whether the expansion in farm productivity in West Bengal

owes more to private investment than government land reforms and extension programs

(e.g., Harriss (1993), Moitra and Das (2004)).

It is possible, however, that part of the expansion in private irrigation investments may

have been stimulated by the tenancy reform. The tenancy reform may have increased the

incentive of tenants to apply water in their farms, and thus shifted the demand function

for groundwater outwards. This in turn could have stimulated investments in tubewells

by water sellers. The process may have been reinforced by increased access to credit for

registered tenants in the early 1980s, for which there are several anecdotal accounts. Since

investments in tubewells are lumpy and involve significant fixed costs, the net effect may

30These have a depth of between 200-500 feet, and cost approximately Rs 200,000. These are usually

provided by the government.31These numbers are calculated for the sample villages on the basis of a 2004 direct household survey of

their landholding histories that we have conducted. For each village a random sample of approximately 20

households per village, stratified by landholding, was selected. The reported areas irrigated average across

all the households in the sample. These numbers correspond closely with those reported by the West Bengal

Census of Minor Irrigation for 1993-94, which reports a 340% rise in shallow tubewells, 161% in dugwells and

45% in deep tubewells in the state between 1987 and 1994. See also Moitra and Das (2004) for a detailed

account of groundwater investments in West Bengal.

27

have been to lower the marginal cost, and hence price, of groundwater.32

In ongoing work with Neha Kumar, we are currently trying to find evidence of such

linkages of tenancy reform and tubewell investments. Our preliminary findings do indeed

find evidence that tenancy reforms and minikit distribution programs had a significant effect

on lowering the cost of irrigation. We expect to report these findings in a subsequent paper.

5 Concluding Comments

To summarize, we have found significant effects of Operation Barga on rice yields and

farm value added per acre, somewhat smaller in magnitude compared with Banerjee et

al, using data from an independent source at a disaggregated farm level, with controls for

endogeneity of program implementation and other concurrent farm extension programs. The

quantitative magnitude of these effects were small compared to those of agricultural kits,

credit and local irrigation facilities delivered by local governments. The effectiveness and

significance of the latter in particular is striking, indicating the importance of public supplies

of agricultural support services in sustaining productivity growth in developing countries.

The importance of their role may arise from a combination of factors: credit constraints

that restrict the ability of most farmers to purchase these inputs on the market, pecuniary

externalities or spillovers in learning associated with inputs that are complementary to

adoption of high-yielding rice varieties, and the public good character of local road and

irrigation programs.

We found evidence that the program raised yields on tenant farms relative to non-tenant

farms, which indicate their role in reducing sharecropping distortions. But this evidence

was statistically less reliable, owing perhaps to the low incidence of leasing. The point

32This is analogous to the effect of expanding market size in stimulating investments in R&D, as modelled

by Dasgupta and Stiglitz (1980) for instance. Suppose there is a Cournot oligopoly of water sellers, who

first select a level of investment in fixed costs, where higher investments lower the marginal costs of supply.

At a second stage, they compete in selling water and earn a margin above marginal costs in equilibrium.

A rightward shift in the demand curve for water will stimulate greater investment, and lower the marginal

cost and price of water at the second stage.

28

estimates imply that the impact of the program on growth of farm yields was quantitatively

much smaller than the impact of farm input supply programs. This owed to the small scale

of the program, related in turn to the low incidence of leasing.

On the other hand, our results support Banerjee et al with regard to the benign impact

of tenancy reform on farm productivity, though we offer a different and more detailed ac-

count of the channels of these impacts. Our quantitative estimate of the effect is somewhat

smaller than theirs, for two reasons. One, they were predicting yields at a much higher level

of aggregation, therefore included effects of the reforms on composition of farms between

different size categories and tenurial status. Second, they did not control for many other

programs administered by local governments that were correlated with implementation of

Operation Barga. Nevertheless, the measured impact is still statistically significant and

robust to alternative data sources, specifications, controls and treatment of possible en-

dogeneity concerns. The low impact of the program relative to other infrastructural and

extension services provided by local governments owed partly to the low incidence of tenancy

in West Bengal villages.

Our interpretation of the effects of the tenancy reform also differs considerably from

earlier literature, which have focused on Marshall-Mill incentive effects alone. We were

surprised by the large spillover effects of the reforms to non-tenant farms, reflecting general

equilibrium or other spillover effects. Identification of the precise channel of such spillovers

is a challenging task: we did not find evidence consistent with a variety of possible channels

in the paper. Our ongoing work suggests that induced effects on private investments in

groundwater irrigation formed one channel by which these reforms caused productivity

improvements to spread to non-tenant farms.

References

Appu P.S. (1996), Land Reforms in India, Delhi: Vikas Publishing House.

Banerjee A., P. Gertler, and M. Ghatak (2002), Empowerment and Efficiency: Tenancy

Reform in West Bengal, Journal of Political Economy, 110(2), 239-280.

29

Bardhan P. (1973), “Size, Productivity and Returns to Scale: An Analysis of Farm-Level

Data in Indian Agriculture,” Journal of Political Economy, 81(6), 1370–86.

—————- (1984), Land, Labor and Rural Poverty (Delhi: Oxford University Press; New

York: Columbia University Press).

Bardhan P. and D. Mookherjee (2000), “Capture and Governance at Local and National

Levels,” American Economic Review, Papers and Proceedings, May 2000, 135-139.

—————– (2004a), “Poverty Alleviation Effort of West Bengal Panchayats,” Economic

and Political Weekly, February 28 2004.

—————– (2004b), “Ideology versus Electoral Competition in Redistributive Politics:

Land Reform in West Bengal,” working paper, Boston University. Title of 2008 revision:

“Determinants of Redistributive Politics: An Empirical Analysis of Land Reforms in West

Bengal, India.”

—————— (2006), “Pro-Poor Targeting and Accountability of Local Governments in

West Bengal,” Journal of Development Economics, 79, 303-327.

Bardhan P. and C. Udry, Development Microeconomics, Oxford University Press, Oxford,

1999.

Bell C. (1977), ”Alternative Theories of Sharecropping: Some Tests Using Evidence from

Northeast India”, Journal of Development Studies, 13 (4), 317-46.

Binswanger H., Deininger K., and Feder G. (1993), “Power, Distortions, Revolt and Reform

in Agricultural Land Relations,” in J. Behrman and T.N. Srinivasan (Ed.), Handbook of

Development Economics, vol. III, Amsterdam: Elsevier.

Besley T. and R.Burgess (2000), “Land Reform, Poverty Reduction, and Growth: Evidence

From India”, Quarterly Journal of Economics, May, 389–430.

Boyce J. (1987), Agrarian Impasse in Bengal: Institutional Constraints to Technological

Change, Oxford: Oxford University Press.

30

Braido L. (2006), ‘Evidence on the Incentive Properties of Share Contracts,’ mimeo, Getulio

Vargas Foundation, Rio de Janeiro.

Chiappori P.A. and B. Salanie (2001), ”Testing Contract Theory: A Survey of Some Recent

Work” , in Advances in Economics and Econometrics, vol 1, M. Dewatripont, L. Hansen

and S. Turnovsky eds, Cambridge University Press.

Dasgupta P. and J. Stiglitz (1980), “ ”, Economic Journal.

Datta Ray S. (1994), “Growth of Agriculture in West Bengal,” Economic and Political

Weekly, 29(47), 19 Nov 1994, 2999-3000.

Dutta, B., D. Ray and K. Sengupta (1989), “Contracts with Eviction in Infinitely Repeated

Principal Agent Relationships,” in Pranab Bardhan (ed.) The Economic Theory of Agrarian

Institutions, Clarendon Press, Oxford.

Eswaran M. and A. Kotwal (1986), “Access to Capital and Agrarian Production Organiza-

tion,” Economic Journal.

Grossman G. and E. Helpman, “Electoral Competition and Special Interest Politics,” Re-

view of Economic Studies, 63, 265-286, 1996.

Harriss J. (1993), “What is Happening in Rural West Bengal? Agrarian Reform, Growth

and Distribution,” Economic and Political Weekly, June 12 1993, 1237–1247.

Honore B. (1992), ‘Trimmed LAD and Least Squares Estimation of Truncated and Censored

Regression Models with Fixed Effects,’ Econometrica 60(3), 533-566.

Johnson D.G. (1950), ”Resource Allocation under Share Contracts”, Journal of Political

Economy, 58(2), 111-23.

Lieten G.K. (1992), Continuity and Change in Rural West Bengal. Sage, New Delhi and

London.

Moitra B. and Das P.K. (2004), ‘Groundwater Markets in West Bengal, India: Emer-

gence, Evolution and Market Structure’, SANEI V Report, October 2004. Available at

http://www.jueconomics.in/faculty/papers/.

31

Mookherjee D. (1997), ‘Informational Rents and Property Rights in Land,’ in Property

Relations, Incentives and Welfare, J. Roemer (ed.), MacMillan Press.

Sengupta S. and H. Gazdar (1996), “Agrarian Politics and Rural Development in West Ben-

gal,” in Dreze J. and A. Sen, Editors, Indian Development: Selected Regional Perspectives,

Oxford India Paperbacks, Oxford University Press, New Delhi.

Shaban R. (1987), ‘Testing Between Competing Models of Sharecropping,’ Journal of Po-

litical Economy, 95(5), 893–920.

Singh N. (1989), ”Theories of Sharecropping”, in P. Bardhan (ed.), The Economic Theory

of Agrarian Institutions, Clarendon Press, Oxford.

32

33

TABLE 1: VILLAGE CHARACTERISTICS

IN SAMPLE VILLAGES, 1978 AND 1998

1978 1998

Number of households 228 398

Operational land-household ratio (acre/hh) 1.54 0.87

% households landless 47.3 52.3

% households marginal (0–2.5 acres) 35.2 39.1

% households small (2.5–5 acres) 11.2 6.4

% households medium (5–12.5 acres) 4.7 2.0

% households big (12.5– acres) 1.6 0.3

% land small 56.7 73.9

% land medium 23.9 18.5

% land big 19.5 7.6

% poor households low caste 38.3 39.8

% upto small households illiterate 44.1 31.9

% big households illiterate 4.4 3.2

% households in nonagricultural occupation 41.1 51.4

Population-Bank ratio 41.6 23.1

‘Poor’ household is either landless or marginal landowner

‘Upto small’ household is either landless, marginal or small landowner

All land information pertains to distribution of cultivable non-patta land owned

Source: indirect household survey;

Population-bank ratio from West Bengal Economic Review, various years

TABLE 2: LAND REFORMS

1978 Average 1998 Average

% land area appropriated 16.4* 15.3

% land area, titles distributed 1.4 5.4

% households receiving titles 4.9 14.9

% land area, tenancy registration 2.4 6.1

% households registered 3.1 4.4

% tenants registered 43.4 51.2

Average across sample villages, weighted by operational land areas

Source: Block Land Records Offices for land reforms implemented

Indirect household survey, for distribution of operational land and tenancy)

*Only available for 34 villages

TABLE 3: TRENDS IN PUBLIC SUPPLIES OF AGRI. INPUTS

1982 1985 1990 1995

Minikits per household 0.11 0.11 0.08 0.06

IRDPa Credit per household 36 29 25 18

Loc. Govt. Irrigation Expenditureb 5233 4265 1485 2627

Loc. Govt. Road Expenditurec 6470d 4501 2501 4572

Loc. Govt. Employment Mandays per household 3.9 3.0 2.2 1.9

Area Irrigated by State Canals (hectares) 72793 72168 79774 84672

State Road Length (Km) 1271 1282 1309 1320

Cumulative % land area, titles distributed 4.5 5.8 6.2 6.3

Cumulative % land area with tenancy registration 1.9 4.2 4.9 5.1

Rows 1–5: Average yearly flow in sample villages, wtd. by land areas

Rows 6–7: stocks at district level, West Bengal government data

a: IRDP Credit Subsidy, 1980 prices;

b,c: Expenditure out of Employment Program Funds, 1980 prices; d: for year 1983

Source: Block Agricultural Dev. Offices, Lead Banks,

Gram Panchayat budgets, West Bengal Economic Review

34

TABLE 4: CROPPING PATTERNS AND RETURNS

Percent of Total Value Added

Total Cropped Area per acre

Area Devoted (1980 prices)

Rice (HYV) 27.5 4137

Rice (non-HYV) 40.0 1925

Potato 3.2 6831

Jute 11.9 3497

Wheat 4.6 1347

Pulses and Vegetables 5.1 2172

Oilseeds 5.9 1808

Tobacco 1.4 3266

Simple Average across sample villages

Source: Cost of Cultivation Surveys

TABLE 5: TRENDS IN FARM PRODUCTIVITY, INCOMES, WAGES

1982 1985 1986 1990 1991 1995

Cropped Area (acres)∗ .99 1.07 .97 1.06 .98 .98

% Rice Area HYV 1.24 7.63 16.54 32.92 45.87 52.70

Rice Value Added per acre 1055 1565 1624 2983 4135 5546

Value Added per acre 703 854 913 1272 1293 1454

Value Added per farm 7613 10915 11223 19070 20553 27349

Hired Labor Wage Rate .57 .54 .83 .76 .81 .99

Hired Labor Annual Hrs/Acre 180 188 242 279 319 265

Averaged across farms, weighted by cropped areas, excepting ∗

All rupee figures deflated by cost of living index, 1974=100

Source: Cost of Cultivation Surveys

35

36

TABLE 6: TENANCY TRENDS

YEAR % FARMS LEASING % AREA LEASED

1982 2.13 12.98

1983 3.24 10.34

1984 4.41 13.11

1985 3.38 6.94

1986 0.44 1.2

1987 0.44 1.14

1988 1.02 2.26

1989 0.73 0.89

1990 0.43 2.07

1991 1.17 6.54

1992 2.5 10.33

1993 2.35 3.86

1994 2.41 6.31

1995 1.98 4.27

37

TABLE 7: EFFECT OF INTERVENTIONS ON FARM PRODUCTIVITY

(OLS ESTIMATES)

(1) (2) (3) (4)

Tenant Dummy -0.049 - -.053 -

(.065) (.056)

Tenant Dummy*% Land Registered 0.125 - 0.253 -

(.186) (.157)

% Years Tenant*%Land Registered - 22.635 - 26.641

(19.272) (19.547)

% Land Registered - - 0.423*** 0.423***

(.128) (.128)

% Land Titled - - 0.187 0.187

(.121) (.120)

Minikits per household - - 0.494*** 0.496***

(.167) (.167)

IRDP Credit per household - - 0.001** 0.001**

(.000) (.000)

Empl. Mandays per household - - 0.048 0.048

(.031) (.031)

Log Local Govt. Irrigation Exp. - - 0.040** 0.040**

(.012) (.012)

Log Local Govt. Road Exp. - - -0.015 -0.015

(.010) (.010)

Controls A A A,B A,B

Farm, Year Dummies Y Y Y Y

Number obs., farms, w-R sq. 2438,631,.08 2438,631,.08 2085,539,.20 2085,539,.20

Dependent variable: Log Farm Value Added per acre

A Controls: Farm acreage and square

B Controls: Rice price, annual rainfall (village); state canals, roads (district level)

All interventions cumulated at village level until previous year

Robust standard errors in parentheses, clustered at village level

***,**,* denote significant at 1,5,10% respectively

38

TABLE 8: EFFECT OF INTERVENTIONS ON FARM PRODUCTIVITY

(IV ESTIMATES)

Full Sample Owner-Cultivators Owner-Cultivators

Only Only

IV OLS IV

Tenant Dummy -0.034 - -

(.058)

Tenant Dummy*% Land Registered 0.224 - -

(.182)

% Land Registered 0.901*** 0.432*** 0.885***

(.238) (.131) (.257)

% Land Titled 0.104 0.178 0.225

(.163) (.148) (.503)

Minikits per household 0.890*** 0.505*** 0.878***

(.239) (.173) (.295)

IRDP Credit per household 0.001*** 0.001** 0.001***

(.000) (.000) (.000)

Empl. Mandays per household 0.111** - -

(.05)

Log Local Govt. Irrigation Exp. 0.036* 0.039*** 0.034*

(.019) (.012) (.018)

Log Local Govt. Road Exp. -0.026** -0.016 -0.025*

(.010) (.010) (.010)

Controls A,B A,B A,B

Farm, Year Dummies Y Y Y

Number obs., farms 2075,534 1993,520 1981,516

Instrumented: %Land Registered, %Land Titles, Minikits and IRDP per hh

First stage F values 2080.4, 26.16, 60.63, 23.71;

p-values 0.00,0.00,0.00,0.00; partial R-sq. .95,.48,.62,.36 respectively

Dependent variable: Log Farm Value Added per acre

A Controls: Farm acreage and square

B Controls: Rice price, annual rainfall (village); state canals, roads (district level)

All interventions cumulated at village level until previous year

Robust standard errors in parentheses, clustered at village level

***,**,* denote significant at 1,5,10% respectively

39

TABLE 9: EFFECT OF INTERVENTIONS ON VILLAGE PRODUCTIVITY

OLS IV

%Marginal Farms Leasing*% Land Registered -0.605 1.801

(5.341) (5.227)

%Medium Farms Leasing*% Land Registered 1.922* 1.829

(1.066) (1.526)

%Large Farms Leasing*% Land Registered -7.286 -5.674

(10.230) (10.850)

% Land Registered 0.385*** 0.683***

(.135) (.226)

% Land Titled 0.069 0.018

(.111) (.169)

Minikits per household 0.397** 0.636***

(.173) (.273)

IRDP Credit per household 0.000 0.001

(.000) (.000)

Empl. Mandays per household 0.043*** 0.093***

(.026) (.047)

Log Local Govt. Irrigation Exp. 0.039*** 0.035**

(.013) (.017)

Log Local Govt. Road Exp. -0.012 -0.020

(.012) (.012)

Controls B,C B,C

Farm, Year Dummies Y Y

Number obs., villages 261,67 261,67

Marginal Farm: <2.5 acres; Medium Farm: between 2.5 and 5 acres; rest Large Farms

Instrumented in Col. 2: %Land Registered, %Land Titles, Minikits and IRDP per hh

Dependent variable: area wtd. average of log farm value added per acre in each village-year

B Controls: Rice price, annual rainfall (village); state canals, roads (district level)

C Controls: % village area distribution between size classes, and % of each size class leasing in

All interventions cumulated at village level until previous year

Robust standard errors in parentheses, clustered at village level

***,**,* denote significant at 1,5,10% respectively

40

TABLE 10A: IMPLIED EFFECTS ON FARM PRODUCTIVITY

FOR AVERAGE VILLAGE

WTD MEAN WTD MEAN IV COEFF PRED. IMPACT

1982 1995 VA/ACRE

% Land Registered .019 .051 .714 +5%

Minikits/HH .11 1.26 .694 +528%

IRDP Credit/HH 36 359 .0012 +144%

Loc. Govt. Irrig. Exp. 4.04 4.74 0.037 +6%

Predicted change of value added per acre implied by estimated coefficients

resulting in a hypothetical village from change in concerned program

equal to weighted average of observed changes in the entire sample between 1982-95

TABLE 10B: AVERAGE PREDICTED EFFECT (% CHANGE) IN

FARM PRODUCTIVITY, DIFFERENT PERIODS

1981-85 1986-90 1991-1995

% Land Registered .18 .81 .89

Minikits/HH 428.34 79.64 8.58

IRDP Credit/HH 82.56 17.72 9.91

Loc. Govt. Irrig. Exp. 74.21 3.86 1.39

Average of predicted percent change of value added per acre in sample villages implied by

estimated coefficients and actual changes observed in those villages;

Operational land areas used as weights

41

TABLE 11: COMPARING INPUT ALLOCATIONS IN RICE CULTIVATION

BETWEEN LEASED AND OWNED PLOTS

HYV Rice non-HYV Rice

Leased Dummy -.21*** -.22 -.06 -.08

(.08) (.14) (.08) (.17)

Leased Dummy*% Land Registered .12 -2.52

(.10) (3.16)

No. obs., farms 1828,576 1590,496 2234,660 1911,572

Within-R sq. .031 .037 .06 .08

OLS Regressions; Dependent Variable: Log of Input Expenditures/Value of Output

All regressions include farm and year dummies

Robust standard errors in parentheses, clustered at village level

***,**,* denote significant at 1,5,10% respectively

42

TABLE 12: LAND REFORM AND PUBLICLY SUPPLIED INPUTS:

DIFFERENTIAL EFFECTS ON SMALL, MARGINAL FARMS

Log Farm Value Added Log Farm Value Added

Per Acre (IV) Per Acre (IV)

% Land Registered 0.694** 0.681***

(.263) (.251)

% Land Registered*Small Farm 0.029

(.026)

% Land Registered*Marginal Farm -0.021

0.019)

% Land Titled 0.592 0.542

(0.631) (0.437)

% Land Titled*Small Farm -0.097

(0.379)

% Land Titled*Marginal Farm -0.085

(0.189)

Minikits/HH 0.562** 0.608**

(0.250) (0.251)

Minikits/HH*Small Farm 0.036

(0.028)

Minikits/HH*Marginal Farm -0.003

(0.026)

IRDP Credit/HH 0.001*** 0.001**

(0.00) (0.001)

IRDP Credit/HH*Small Farm -0.00

(0.00)

IRDP Credit/HH*Marginal Farm 0.00

(0.00)

43

TABLE 12 continued

Log Farm Value Added Log Farm Value Added

Per Acre (IV) Per Acre (IV)

(Log) Loc. Govt. Irrigation Exp. 0.029* 0.029**

(0.014) (0.015)

LG Irrig Exp*Small Farm -0.011

(0.008)

LG Irrig Exp*Marginal Farm -0.016*

(0.009)

(Log) LG Road Exp. -0.023 0.019*

(0.014) (0.011)

LG Road Exp*Small Farm 0.000

(0.011)

LG Road Exp*Marginal Farm -0.003

(0.008)

LG Empl. Mandays/HH 0.208** 0.125**

(0.081) (0.052)

Mandays/HH* Small Farm -0.108

(0.081)

Mandays/HH*Marginal Farm -0.002

(0.044)

No. Obs., farms, Within-R sq. 2091,542, .16 2091,542, .17

Small: < 5 acres; Marginal:< 2.5 acres

Controls: farm,year dummies; A,B (Table 8)

IV regression, details as in Table 8

Robust standard errors in parentheses, clustered at village level

***,**,* denote significant at 1,5,10% respectively

44

TABLE 13: IMPACT OF INTERVENTIONS ON CROPPING PATTERNS

TOTAL HYV RICE LOCAL RICE POTATO

AREA AREA AREA AREA

IV IV IV IV

% Land Registered 1.245** 0.738* 0.765** -0.226**

(0.523) (0.382) (0.366) (0.110)

% Land Titled 0.225 0.465 -0.622 0.045

(0.678) (0.466) (0.695) (.196)

Minikits/HH 0.830 0.748** .125 .264*

(0.522) (0.349) (.455) (.155)

IRDP Credit/HH 6.2e-4 1.7e-4 1.2e-4 -1.9e-4

(6.7e-4) (5.8e-4) (7.5e-4) (1.8e-4)

Log Rice Price 0.257 0.175* -0.056 -0.057

(real)b (0.332) (0.162) (.104) (.046)

Number obs., farms 2071, 539 2099,542 2099, 542 2099, 542

Within-R sq. .05 .04 .03 .07

b: Deflated using regional CPI for Agricultural Workers

Other controls in previous tables included

Robust standard errors in parentheses, clustered at village level

***,**,* denote significant at 1,5,10% respectively

45

TABLE 14: IRRIGATION EXPANSION, BY SOURCE

Year Canals Deep Tubewells River lift/pond Shallow Tubewell Others

1981 5.20 2.81 7.5 7.62 0.45

1982 5.20 2.81 7.45 8.79 0.46

1983 5.21 2.92 7.54 9.03 0.46

1984 5.22 3.08 8.17 10.97 0.46

1985 5.22 3.22 8.67 13.18 0.46

1986 5.27 3.30 9.08 14.09 0.48

1987 5.28 3.42 9.32 14.81 0.48

1988 5.35 3.55 9.43 15.56 0.54

1989 5.4 3.66 9.97 17.01 0.54

1990 5.41 4.05 11.9 19.43 0.55

1991 5.42 4.10 12.09 20.85 0.55

1992 5.45 4.3 12.43 23.22 0.55

1993 5.45 4.35 12.52 24.27 0.56

1994 5.47 4.41 13.44 29.29 0.68

1995 5.82 4.51 14.85 31.39 0.68

Average fraction of cultivable area in sample villages irrigated, by source

Source: Household Land Survey Data